AI Model Predicts Vocational Graduate Job Paths with 82% Accuracy

A new machine learning system is helping predict where vocational college graduates end up after they leave school. Researchers built a smart computer model that looks at student data to guess whether someone will get a job right away, go back to study more, or struggle to find work.

The model was trained using information from 2,270 graduates from a Chinese vocational college. These students graduated between 2016 and 2019. The computer looked at 27 different pieces of information about each student, including their grades, test scores, and activities they did outside of class.

Most students in the study found jobs easily - about 79% got work right away. Around 18% went back to school, while only about 3% couldn't find work. The researchers used a special technique called TPE to fine-tune the model's settings, making it better at guessing correctly.

When tested on new data, the model got 82% of its predictions right overall. However, it had a harder time with the smaller groups - especially predicting who would go back to study or who might be unemployed. The model beat seven other common prediction methods in head-to-head tests.

The researchers also used a tool called SHAP to figure out which student information mattered most. They found that grades from multiple semesters, entrance exam scores, and hours spent volunteering were the biggest clues about where a student would end up. This suggests that both academic performance and community involvement play important roles in shaping graduate outcomes.